When you break the one non-negotiable rule of AI: “keep humans in the loop”, bad things happen.
A friend sent me the link to an article on the 1Password blog today about the recently released MoltBot AI Agent, which quite a few people are experimenting with on their Macs.
1Password, as a credential management system, focused its article on the need to rethink agent security.
“The mistake the industry is making right now is treating agent security like normal app security. A familiar consent screen. A one-time approval. A set of scopes. Then we assume the future behavior will match the intent of that one moment.
That model breaks the second you hand autonomy to something that is adaptive and non-deterministic by design. The agent changes. The tasks change. The context changes. The approval you gave last week is used in new and unexpected ways today.”
I completely agree with this statement; however, we have a much bigger problem with systems like MoltBot.
MoltBot exhibits “dynamic behaviors born out of an agentic loop that takes a goal and improvises a plan, grabbing whatever tools it needs to execute. It can apply general world knowledge, specific skills, and near-perfect memory into organized action toward objectives you set, and, more sobering, objectives it decides to set for itself.”
Seriously?
Somebody thought it was a good idea to give an AI system autonomy to set its own guardrails and acquire new skills that it wasn’t programmed for?!?! They have broken the one totally non-negotiable rule of AI and Intelligent Automation. ALWAYS keep the human in the loop and in control.
What really concerns me is that this capability was not confined to a research lab with strict protocols for ethical AI, but was released into the wild for enthusiasts to play with, disregarding the very strong potential for misuse (due to incompetence or ill intent).
This is where I feel the necessity to reiterate the importance of responsible and ethical use of AI – not just by big companies, but by every individual who uses it. If you don’t know what I’m talking about, do a quick search and watch a couple of YouTube videos on the subject.
AI and Intelligent Automation are powerful tools with tremendous potential to increase human productivity in ways we have never seen before. But these tools can also be harmful if not used responsibly. Before you install MoltBot or any similar tool, or build autonomous AI agents of your own, please give serious thought to the impact of your actions and the steps that you need to take to keep the “human in the loop” and maintain control of your creation.
